Deformation parameter estimation on IC failure locations in liquid crystal images

Abstract

This paper focuses on application for non-rigid contour tracking in heavily cluttered liquid crystal images. Based on the properties of non-rigid contour movements, cascading image sequences are conducted for contour motion vector estimation using a hexagon-diamond search. This enables retrieving the motion for the failure parts in the crystal images and accurately tracks them while the shape changes on the object. Experimental results show that the hexagon-diamond search successfully tracks the non-rigid movements.